This paper examines how different uncertainty measures affect the unemployment level, inflow, and outflow in the U.S. across all states of the business cycle. We employ linear and nonlinear causality-in-quantile tests to capture a complete picture of the effect of uncertainty on U.S. unemployment. To verify whether there are any common effects across different uncertainty measures, we use monthly data on four uncertainty measures and on U.S. unemployment from January 1997 to August 2018. Our results corroborate the general predictions from a search and matching framework of how uncertainty affects unemployment and its flows. Fluctuations in uncertainty generate increases (upper-quantile changes) in the unemployment level and in the inflow. Conversely, shocks to uncertainty have a negative impact on U.S. unemployment outflow. Therefore, the effect of uncertainty is asymmetric depending on the states (quantiles) of U.S. unemployment and on the adopted unemployment measure. Our findings suggest state-contingent policies to stabilize the unemployment level when large uncertainty shocks occur.